Retention rate

Analytics tools report it two ways. A single percentage, often labelled average percentage viewed, says how much of the video a typical viewer saw. A retention graph plots what share of viewers remained at every second, which is the more useful view because it shows exactly where people left rather than how many did overall.

It matters because recommendation feeds are built on it. A clip that holds most of its viewers through to the end tells the ranking system that the opening promised the right thing and the body delivered, so it gets shown to a wider audience. For a clipper or a podcaster the graph is the fastest feedback available on which moments and which openings actually work.

The nuance is that the shape tells you more than the number. A steep drop in the first two seconds is a hook problem, a steady slide is a pacing problem, and a dip at one point is a specific moment worth cutting. Short clips also run naturally higher than long videos, and loops can push the figure past one hundred percent, so the number is only comparable between clips of similar length on the same platform.

Practically, retention improves by starting the cut inside the moment, trimming anything that explains instead of shows, and placing a change on screen just before the point where the graph usually dips. Tools that score segments of a long recording for standalone interest are essentially predicting retention before the clip is posted.